ABSTRACT: Research Project I ? Innovative Methods for Constructing Just-in-Time Adaptive Interventions The long-term goal of this project is to improve public health by facilitating the evidence-based construction of effective, individualized, mobile drug abuse prevention and intervention services. This project develops data analytic methods that will enable drug abuse prevention and intervention scientists to more effectively adapt services to individuals' changing needs over time and to more effectively expand the reach of their services. The end result of this project is that these smartphone-based interventions will not only be accessible around the clock (available whenever, wherever, and for as long as needed), but also highly responsive to dynamically changing individual needs. Just-in-time adaptive interventions are composed of operationalized decision rules that input dynamic individual information (e.g., current craving, geographical location, substance use) and output, via the mobile device, strategies (e.g., motivational, cognitive, behavioral, social) that support the individual's long term goal to avoid high-risk behavior. To develop interventions that respond nimbly to dynamically changing individual needs, the intervention designer must address nuanced questions such as, ?At what risk level should the mobile device prompt the individual and recommend the use of a particular service?? and ?Should the length of time an individual has been abstinent and/or the current location of the individual be used in addition to the current risk level to determine whether to recommend the use of a recovery strategy?? Current behavioral theories are, for the most part, silent on these kinds of questions. The analysis of existing intensive longitudinal intervention data could be very informative, but appropriate data analytic methods do not exist. This state of affairs is preventing mobile interventions from fulfilling their potential to provide effective services for reducing drug abuse and HIV. The overarching goal of this project is to integrate ideas from statistics, computer science, and behavioral science to develop data analytic methods that will (1) enable scientists to construct more effective mobile interventions for delivery of drug abuse/HIV prevention and recovery services and (2) inform the development of more dynamic and nuanced behavioral theories.